Every year, many game titles are introduced to the market in response to the great interest for new and exciting games. Due to the great number of available game titles, as well as the high rate of new releases, a game playing enthusiast may not be aware of all the game titles that are available. This may result in the player missing game titles that he would have been very interested in buying or renting due to the nature and contents of the game. It is often difficult for the game manufacturers to identify potential buyers who may be interested in the specific games the manufactures produce. This lack of communication with the buyer results in lost profits, and the manufacturer is left with inefficient ways to introduce and sell a new line of products.
Recommender systems are generally known as technology for allowing automated systems to recommend products or services to users. A typical recommender system provides a specific type of information filtering (IF) technique that attempts to present information items (e.g., movies, music, books, news, images, web pages, etc.) that are likely to be of interest to a user. Such a system helps the user deal with information overload while also providing an opportunity for a seller to make an additional, targeted sale. Such systems have long been employed as part of many electronic commerce websites to generate competitive advantage. Examples of recommender systems currently in use include among others:
Amazon.com (online retailer, includes product recommendations)
Arnie Street (music service)
Baynote (recommendation web service)
Collarity (media recommendation platform)
Daily Me (news recommendation system (hypothetical))
Genius (iTunes) (music service that is part of the iTunes Store)
inSuggest (recommendation engine)
iLike (music service)
Last.fm (music service)
Strands (developer of social recommendation technologies)
Netflix (DVD rental and online movie playing service)
Pandora (music service)
Reddit (news recommendation system)
Slacker (music service)
StumbleUpon (web discovery service)
StyleFeeder (personalized shopping search)
Typically, a recommender system compares a profile associated with a user to some reference characteristics, and seeks to predict the ‘rating’ that the user would give to an item he has not yet considered. For example, visitors to websites such as Amazon.com likely have seen the following type of message upon logging on to the website: “Latest authors you may like including . . . ” followed by a list or graphical presentation of books or other items the website recommends to the user. Similarly, the Netflix website commonly recommends movies to its customers. Such recommendations are typically generated based on the particular customer's past purchases—i.e., items that are “like” items the customer has ordered in the past but without recommending the same exact items the customer has already purchased. Such recommendations can provide customers with useful information to guide them through a wide array of new products. The recommendations can also help website operators make additional sales. Thus, recommendations are often “win-win” for both the customer and the website operator.
Information about specific customers is especially helpful in recommending products to those customers. This is relevant to online or offline purchasing where personalized customer preferences are employed to market additional products and services to prior customers. In an example online operation, a customer preference profile is sometimes obtained by asking the customer to respond to a list of questions related to demographic personal information, buying preferences, and/or from tracking the customer's online activities. A computer program may then be used to model a profile for the client and predict the outcome of the customer's future behavior.
It would be beneficial to apply recommender technology to the marketing of video and computer games, and to provide methods and systems that make available in real time the preferences and buying habits of the computer game enthusiast and tailor these features to make suggestions for future use. However, there are significant challenges (e.g., privacy concerns, parental controls, and other factors) that have in the past stood in the way of launching an effective, efficient recommender system for video game consoles.
The exemplary illustrative non-limiting technology herein solves such problems by providing methods and systems for using information related to a computer game player to provide suggestions about game title availability and other game related features specific to the game player, while providing parental controls, privacy protection and other desirable features.
In one exemplary illustrative non-limiting implementation, an online recommendation service for computer game players is provided. A record of the online activities is generated and maintained for each player. The record includes identities of game titles downloaded, purchased, or gifted, and time spent playing each game by the player. Based on the record, the service provides recommendations of new games to be purchased or rented to the player. In addition the service offers information such as popularity of game titles with other players, popular search strategies and other game related information.
In one exemplary illustrative non-limiting implementation, the online service provides parental control settings for the online sessions, and provides customized settings depending on the playing level/expertise as well as the genre of the game. In addition, the exemplary illustrative non-limiting service offers privacy protection for the player before information is disclosed to generate and maintain the user's online activities. For example, the player can be asked to “opt in” to a game recommender system before any private or usage information about the player's game playing habits or preferences are disclosed to the recommender system. In some non-limiting implementations, there can be a variety of different matching algorithms for recommending with different algorithms being selected on a player-by-player basis depending on past success rates.
Some illustrative non-limiting techniques herein provide weighting of data by age. For example, when a given piece of data is collected from a user, a collection date can associated with that data. This enables the exemplary illustrative non-limiting system to calculate how old the data is. In one exemplary illustrative non-limiting example, older data has less influence and newer data has more influence. This exemplary functionality for example takes into account a game player's changing tastes.
Additional exemplary illustrative non-limiting features and advantages include:                Collect and use how much time a player has played a game as a factor in recommending        Keep track of game titles downloaded, game titles purchased on storage media, game titles gifted, game titles reviewed        Allow users to rate games purchased and use such ratings as a factor in developing recommendations.        Recommendations filtered so that games are not recommended that is outside of a parental controls setting        Number of impressions on a portion of a game catalog can be used as a factor in developing recommendations.        Popularity of titles with other people (e.g., friends) can be used as a factor in developing recommendations        Price sensitivity (e.g., player tends to purchase games that cost 500 points or less) can be used as a factor in developing recommendations.        Ability to collect search strategy (keyword, other search strategy)        Ability to link to an affinity or loyalty account        Game Playing level and/or experience can be used as a factor in developing recommendations.        Ability to factor in personal preferences (e.g., a player who purchases only M games will have only M games recommended to him or her)        Ability to recommend for download only games that the destination device has sufficient storage space to store.        Ability to take into account games that other users having recommended to a game player.        Ability to take genre preference (fighting games, adventure games, sports games, etc.) into account in developing a recommendation        Possible to perform some of the analysis in a distributed manner (e.g., can send down recommendation software to a game console)        Ability to recommend based on particular personally owned video game consoles        Can recommend titles to purchase or rent, or can recommend titles to purchase in a retail store        Can make several attempts at weighting or recommending, and select a method that is most efficient or successful (i.e., works the best) for this particular user for use in the future with this particular user.        Privacy protection: the user must opt in before the video game console will disclose personal or personalized information to the recommendation server.        